Infrared Drying Parameter Optimization

Rochester Institute of Technology

RIT Scholar Works

Theses Thesis/Dissertation Collections

1-26-2015

Infrared Drying Parameter Optimization

Follow this and additional works at:http://scholarworks.rit.edu/theses

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please [email protected].

Recommended Citation

Infrared Drying Parameter

Optimization

Matthew R. Jackson

Thesis submitted to the Faculty of the

Rochester Institute of Technology

In partial fulfillment of the requirements for the degree of

Master of Science

in

Industrial Engineering

Thesis Committee

Dr. Denis Cormier

Dr. Brian Thorn

Department of Industrial and Systems Engineering

DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING

KATE GLEASON COLLEGE OF ENGINEERING

ROCHESTER INSTITUTE OF TECHNOLOGY

ROCHESTER, NEW YORK

CERTIFICATE OF APPROVAL

January 26, 2015

M.S. DEGREE THESIS

The M.S. Degree Thesis of Matthew Jackson

has been examined and approved by the

Thesis committee as satisfactory for the

Thesis requirement for the

Master of Science degree

Approved by:

Dr. Denis Cormier, Thesis Advisor

i

Abstract

In recent years, much research has been done to explore direct printing methods,

such as screen and inkjet printing, as alternatives to the traditional lithographic process.

The primary motivation is reduction of the material costs associated with producing

common electronic devices. Much of this research has focused on developing inkjet or

screen paste formulations that can be printed on a variety of substrates, and which have

similar conductivity performance to the materials currently used in the manufacturing of

circuit boards and other electronic devices. Very little research has been done to develop

a process that would use direct printing methods to manufacture electronic devices in

high volumes.

This study focuses on developing and optimizing a drying process for conductive

copper ink in a high volume manufacturing setting. Using an infrared (IR) dryer, it was

determined that conductive copper prints could be dried in seconds or minutes as opposed

to tens of minutes or hours that it would take with other drying devices, such as a vacuum

oven. In addition, this study also identifies significant parameters that can affect the

conductivity of IR dried prints. Using designed experiments and statistical analysis; the

dryer parameters were optimized to produce the best conductivity performance for a

specific ink formulation and substrate combination. It was determined that for an

ethylene glycol, butanol, 1-methoxy 2- propanol ink formulation printed on Kapton, the

optimal drying parameters consisted of a dryer height of 4 inches, a temperature setting

between 190 - 200°C, and a dry time of 50-65 seconds depending on the printed film

thickness as determined by the number of print passes.

It is important to note that these parameters are optimized specifically for the ink

formulation and substrate used in this study. There is still much research that needs to be

done into optimizing the IR dryer for different ink substrate combinations, as well as

developing a control system to ensure that prints continuously dry the same way. In

addition to the repeatability study, experimenting with the feasibility of using single pass

prints with repeatable performance would also be a worthwhile study. A single print pass

will reduce cycle time, and will reduce ink consumption when compared with double

ii Contents

1 Introduction ... 1

2 Literature Review ... 3

3 Methodology ... 14

3.1 Screening Experiments ... 14

3.2 Final Designed Experiment ... 18

4 Results ... 25

4.1 Feasibility of IR Drying ... 25

4.2 Characterization of Vastex Dryer ... 36

4.3 Designed Experiment and Statistical Analysis ... 41

5 Conclusions and Future Work ... 57

6 References ... 59

7 Appendix ... 61

A. Minitab output for double pass four inch dryer height prints ... 61

iii List of Tables

Table 1: Summary of initial experimental space explored ... 17

Table 2: IR dryer parameter used in experiment to characterize heat setting ... 19

Table 3: IR dryer parameters for prints dried at a height of 2 inches ... 22

Table 4: IR dryer parameters for prints dried at a height of 4 inches ... 22

Table 5: Settings for IR dryer and Xenon sintering system used in the final experiment ... 24

Table 6: Thermal conductivity properties of common materials ... 29

Table 7: P-values for 4 inch 2 pass factorial analysis ... 43

Table 8: P-values for 4 inch 2 pass response surface analysis... 46

Table 9: P-values for 3 inch 1 pass factorial analysis ... 51

iv List of Figures

Figure 1: Subtractive lithographic process vs. additive direct printing process [2] --- 2

Figure 2: Piezoelectric drop ejection [1] --- 3

Figure 3: Example of a screen printer --- 4

Figure 4: Example of a vacuum oven --- 5

Figure 5: Example of a tube furnace --- 6

Figure 6: Example of an infrared (IR) dryer --- 6

Figure 7: Novacentrix pulseforge system --- 8

Figure 8: Xenon flash system --- 8

Figure 9: Intrinsiq LAPS-60 laser sintering system --- 9

Figure 10: Afford-A-Flash infrared dryer --- 14

Figure 11: Inkjet print on kapton --- 15

Figure 12: Fishbone diagram to identify sources of within experiment variability --- 17

Figure 13: Vastex IR drying system --- 18

Figure 14: Belt speed settings as a function of time --- 19

Figure 15a,b,c,d: Substrate temperature for glass and kapton at specific dryer heights and heat settings --- 20

Figure 16a,b: Thermal map at the highest heat setting at height of 2 (a) and 4 (b) inches --- 21

Figure 17: Resistance values ranging between 17 and 30Ω were achieved using low dryer heights and short dry times --- 25

Figure 18: Similar resistance values were achieved using high dryer heights, less dryer power and longer dry times --- 26

Figure 19: Example of an over-dried IR print --- 27

Figure 20: IR dried prints preformed slightly better than vacuum oven dried prints --- 27

Figure 21: Hypothesis test comparing the average resistance of IR dried prints to oven dried prints --- 28

Figure 22: Resistance data for prints sintered within 30 minutes of drying --- 30

Figure 23: Resistance data for prints sintered 3 hours after drying --- 30

Figure 24: Hypothesis test comparing the average resistance of prints sintered immediately after drying to prints sintered three hours after drying --- 31

v Figure 26: Oven dried prints made with standard dots per inch (1270) and higher dots per inch

(1693) --- 32

Figure 27: Resistance range for IR dried prints with two different dots per inch settings --- 33

Figure 28: Resistance range for oven dried prints with two different dots per inch settings --- 34

Figure 29: The temperature in the center of the IR dryer is higher since it caused more discoloration --- 35

Figure 30: Resistance data for IR and oven dried prints at 2 inches --- 36

Figure 31: Resistance data for IR dried and oven dried prints at 4 inches --- 36

Figure 32: Hypothesis test for 1 pass prints dried at 2 inches with the new IR dryer and the oven --- 38

Figure 33: Hypothesis test for 1 pass prints dried at 4 inches with the new IR dryer and the oven --- 38

Figure 34: Resistance data for double pass IR and oven dried prints at 4 inches --- 39

Figure 35: Hypothesis test for 2 pass prints dried at 4 inches with new IR dryer and oven --- 40

Figure 36: Factorial analysis main effects plots for 4 inch 2 pass prints --- 42

Figure 37: Factorial analysis interactions plots for 4 inch 2 pass prints --- 42

Figure 38: Factorial analysis residuals plots for 4 inch 2 pass prints --- 44

Figure 39: Response variable vs. fitted values for 4 inch 2 pass prints --- 44

Figure 40: Contour plots for double pass 4 inch dried prints--- 45

Figure 41: Residual plots for response surface analysis of double pass 4 inch dried prints --- 47

Figure 42: Response variable vs. fitted values for 4 inch 2 pass prints --- 47

Figure 43: Sintered and unsintered IR dried two pass print at a height of 4 inches --- 48

Figure 44: Sintered and unsintered IR dried single pass print at a height of 3 inches --- 49

Figure 45: Factorial analysis main effects plots for 3 inch 1 pass prints --- 50

Figure 46: Factorial analysis interactions plots for 3 inch single pass prints --- 50

Figure 47: Factorial analysis residual plots for 3 inch single pass prints --- 52

Figure 48: Response variable vs. fitted values for 3 inch single pass prints --- 52

Figure 49: Contour plots for 3 inch single pass prints --- 53

Figure 50: Response surface residual plots for 3 inch single pass prints --- 54

Figure 51: Response variable vs. fitted values for 3 inch single pass prints --- 55

1

1

Introduction

In the world of electronic devices, controlling the location and amount of material

coated onto a substrate has always been important. Lithographic processes are typically

used to make circuit boards, semiconductors, and other electronic devices. These

processes require several steps, which involve coating a given substrate with a layer of

photosensitive material to be patterned, and then using a mask and chemicals to remove

selected material from the layer. If the resulting pattern is correct, additional chemical

processing transfers the pattern to the layer beneath the photosensitive material. The

photo resist material is then stripped away and the sample cleaned. This process serves

the purpose of controlling where and how much conductive material remains on the

substrate. This process is referred to as a subtractive process, since it involves taking

material away. It results in a significant amount of wasted material, it has pattern design

limitations, and there may be difficulty aligning congruent images if either the mask or

the substrate expands or shrinks during drying [1].

As a result of these challenges with traditional lithographic processes, other coating

techniques have been explored to see if it was possible to develop a method that only

deposited material where it was needed. Printing methods such as inkjet and screen

printing quickly became areas of focus, as these methods would allow those designing

various electronic devices, such as circuit boards and other electronic devices, to only put

material where it was needed, and eliminate several other steps from the traditional

lithographic process. A comparison of the two methods can be found in Figure 1. Direct

printing methods, such as inkjet and screen printing, reduce the waste of material,

chemicals, and photoresist from the etching process. In addition, these methods eliminate

design restrictions and difficulty aligning images. However, inkjet and screen printing

methods require the development of electronic materials in the form of inks and pastes so

that they can be printed onto a given substrate. As a result conductive inks and pastes

have become a major area of research within the last decade. The focus of this work is on

copper inks, since they offer a cheaper alternative to silver based inks, and could be

2 Figure 1: Subtractive lithographic process vs. additive direct printing process [2]

Lithographic Process

3

2

Literature Review

Conductive inks and pastes are much different than just simple metal conductors.

Specifically, inks consist of metal particles, in some cases an organic polymer, and

solvent [3]. Inks are typically made with spherical nano-sized particles, and usually have

a low viscosity so that they can be printed easily. This requires a relatively low solid

loading fraction (the amount of conductive particles) in order to keep the viscosity low.

These inks are usually printed through piezoelectric print heads in an industrial setting.

Piezoelectric print heads contain micro-machined chambers with one or more walls

fabricated from a ceramic, such lead-zirconate-titanate (PZT). The piezoelectric expands

or contracts when a voltage is placed on it. This mechanical deflection creates a volume

change within the chamber and an acoustic wave, which drives a droplet of liquid

through the hole in the nozzle plate of the print head [1]. This process can be seen in

Figure 2.

Figure 2: Piezoelectric drop ejection [1]

Pastes, on the other hand, use mostly micron size particles with a mix of

nano-sized particles in order to maximize packing density. The shape of the micron-nano-sized

+

-a) Voltage flexes membrane expanding volume

4

particles, for example flakes or spheres, can also be varied to increase packing density.

Pastes are typically printed using a screen printer, which can be seen in Figure 3. A

screen printer consists of a screen, which has a designed pattern, and a squeegee that is

controlled automatically. The screen paste is placed directly on the screen, and the

substrate that is to be coated with the paste is placed on a platen underneath the screen.

The squeegee then moves across the screen and forces the paste through the holes in

screen and onto the substrate.

Figure 3: Example of a screen printer

After the inks and pastes are printed, they are usually dried using heat provided by

some type of an oven or a furnace. All materials have basic properties, one of which is

their melting point. These properties are usually cited for bulk materials, which are

processed to obtain a maximum density. Thin films are often less dense than bulk

materials, and therefore may exhibit different properties. For conductive inks, it has been

reported that when the size of the metal particles decreases below 100 nm, their thermal

processing temperature is much lower [4]. As a result there are several different types of

ovens and furnaces that can be used to dry conductive inks and pastes.

Squeege e

Platen for Substrate

5

In the processing of conductive inks, it is important to understand that there is a

difference between drying and sintering. The drying process refers mainly to solvent

removal from the printed pattern. Sintering is where the distance between the particles

decrease to the point where the particles touch and are fused together forming and infinite

network, thus making the ink or paste conductive [2]. For some inks, drying and sintering

can happen in the same step. However, this is not necessarily the case for all inks.

One of the most common drying tools currently being used is a vacuum oven, seen in

Figure 4. Because this tool dries inks and pastes under vacuum, solvent volatilization can

happen at very low temperatures and allows for clean burn-off and evaporation of

solvents. However, if a low drying temperature is used, the drying time will need to be

increased in order to remove all the solvent.

Figure 4: Example of a vacuum oven

Another common drying tool is a tube furnace, which can be seen in Figure 5. A tube

furnace allows users to dry inks and pastes at much higher temperatures than a vacuum

oven. It allows users to program specific drying conditions that control the rate at which

the furnace heats, which can be very important for certain substrates. The tube furnace

can have a controlled atmosphere, such as Argon, as opposed to vacuum or air. Since

tube furnaces are readily available, they are also used to evaporate solvents from

6 Figure 5: Example of a tube furnace

While vacuum ovens and tube furnaces are commonly used in laboratories and

research settings, they may not be practical for high volume manufacturing lines, such as

those used in roll to roll printing. One of the more common techniques used in roll to roll

printing is infrared (IR) drying, which can be seen in Figure 6. This dryer transfers

energy through electromagnetic radiation and can be used in either atmosphere or

vacuum. It is preferred for roll to roll printing processes because it is a noncontact form

of drying. Also, the drying time for conductive inks and pastes is seconds or minutes as

opposed tens of minutes or hours required for vacuum ovens.

Figure 6: Example of an infrared (IR) dryer

During the drying process, the solvent is volatized and removed from the print. This

may result in a reduction in the distance between the metal particles. Depending on

7

sintering process may also happen during the drying step. For example, silver inks and

pastes can be dried and sintered simultaneously. This is due in part to the fact the

nano-size silver is relatively easy to produce, and because the melting point of silver is

significantly reduced when it is less than 100 nm in size, which results in minimal energy

being needed to sinter it [4]. The melting point of bulk silver is typically 960°C.

However, the melting point drops to around 150°C for silver nanoparticles under

approximately 100 nm. Silver can also be sintered and dried simultaneously since silver

oxide is conductive.

Copper inks and pastes, on the other hand, require separate drying and sintering steps.

This is partially due to the fact that copper requires significantly more energy to sinter

than silver does. The melting point of bulk copper is 1085°C, roughly 125°C higher than

that of silver. This difference in melting point holds true for these two materials even

when they are of a size smaller than 100 nm. Furthermore, nano-size copper is difficult to

produce. The biggest reason a separate sintering step is required for copper is because

copper oxide forms when it is handled in air for an extended period of time. This

significantly hinders conductivity [4]. In order to prevent oxidation of the copper

particles, they are usually coated with a polymer. This also helps with dispersion of the

particles in a solvent system. In order for these inks and pastes to become conductive, the

polymer must be removed from the particles so they can fuse together. This happens

during the separate sintering step. One method involves exposing the copper prints to a

quick pulse of UV light, which removes the polymer from the particles, reduces any

oxide on the particles, and fuses them together. The most common pieces of equipment

on the market specifically made for sintering materials with UV light, such as copper, are

the Xenon Flash system and the Novacentrix Pulse Forge system.

Both systems operate on the principle that using shorter duration pulses results in a

higher peak power of energy being delivered to the material. This higher peak power

means that more energy can be put into the material, which enhances its ability to sinter.

8

sintered in air as opposed to an inert environment. This type of sintering step is

compatible with a roll to roll printing operation [4].

In a high volume printing operation, a flash lamp configuration with a conveyor belt,

such as the Novacentrix system shown in Figure 7, is preferred. This Novacentrix system

allows the user to control pulse width, pulse energy, and wavelength. A bench top Xenon

Flash System more suited for laboratory experimentation is shown in Figure 8.

Figure 7: Novacentrix pulseforge system

Figure 8: Xenon flash system

While either the Novacentrix or Xenon systems can be used to sinter both inks and

screen pastes, laser sintering has recently emerged as yet another promising method for

sintering patterns with narrow lines, high metal density, and heat absorbing substrates.

9

together. It is important, when sintering copper with a laser, to move the laser at a speed

slow enough for it to penetrate the material, but rapidly enough so that it sinters the entire

pattern before the copper can oxidize [5]. An example of a laser sintering system can be

seen in Figure 9.

Figure 9: Intrinsiq LAPS-60 laser sintering system

After the sintering step is complete, there are sometimes visible defects in the printed

pattern. Some common defects in both inkjet and screen printed patterns are cracks and

adhesion failures. Cracking occurs due the larger volume reduction that may happen

when prints are being sintered and the protective organic polymer is being removed to

make the structure more dense and conductive. As a result of cracking, the print has poor

adhesion to the substrate, and very poor electrical conductivity [6]. In order to achieve

good electrical conductivity, poor adhesion and cracking must be eliminated. One

possible solution to limiting cracking and getting better substrate adhesion is using

nano-sized glass frit. If a glass frit can be selected such that it melts during sintering and can

form a stable dispersion with the metallic nanoparticles without clogging the print head

nozzles, resistivities of 2.52 * 10-2 Ω-m have been achieved with silver inks [7]. By

comparison, the bulk resistivity of silver is 1.59 *10−8Ω-m.

After learning about the additional steps and equipment needed to sinter copper

materials, one may ask the question “Why use copper when silver is so much easier to work with?” What makes copper attractive for applications where silver is currently

being used, such as Organic Light Emitting Diodes (OLEDs), Photovoltaics, Radio

Frequency Identifiers (RFID’s), etc., is the significant cost savings that comes from using

10

which makes them impractical to use in high volume, low cost applications [8]. This

makes copper an attractive alternative.

While the market for printed electronics has started to materialize, the focus of

research has begun to shift to developing cost effective solutions for printed electronics

applications. Ian Clark, marketing and sales director for Intrinsiq Materials, said

“Although there are few commercial applications for copper inks so far, it is likely that

this will change in the next one to three years as the various development programs

designed to integrate more cost effective copper inks into the manufacturing processes for

electronic devices come to fruition. There is a lot of development activity in the area, not

only from the ink suppliers like ourselves, but also from the print head makers and the

manufacturers of laser and broad band flash curing equipment” [9]. As copper continues

to be an area of focus for researchers, it is important that they take into consideration the

manufacturability of their final product. This means ensuring that their final product will

perform well when using equipment that is well suited for high volume production, such

as a roll to roll printing process. In depth understanding and testing of the product must

be done with equipment that will be used in a high volume printing operation.

A study was done with silver ink, comparing prints on glass that were dried in a

conventional heat oven and an IR dryer. Since silver can be dried and sintered in the

same step, both conventional and IR drying can produce conductive prints. The oven

dried print was dried at 200°C for 5 minutes and produced a sheet resistance of 1.2 Ω/□,

while an identical print was dried under the IR dryer for 3 minutes and produced a similar

sheet resistance of 1.0 Ω/□. The study also provided evidence that the substrate is a

critical parameter in determining drying time, as it significantly affects sintering

efficiency. Two prints, one on glass and one on paper were dried under the IR dryer at a

height of 6 cm for 2 minutes. The glass substrate print heated to 140°C, and had a sheet

resistance of 351.2 Ω/□, while the paper substrate print reached 210°C and had a sheet

resistance of 1.97 Ω/□ [10].

Another study suggested using a hybrid drying process of hot air first, followed by IR

11

were printed on polyethylene terephthalate (PET). It was found that using only IR to dry

these prints resulted in wrinkling of the PET and variation in the sheet resistance values.

Alternatively, drying with only hot air resulted in stable sheet resistances, but required a

long time. In order to try to get the best of both processes, the two were combined, and it

was determined that hot air must be used prior to the IR dryer. When the IR dryer was

used first, it started to sinter the top layer of the print before all the solvent had been

removed, thus leaving solvent trapped in the print and degrading electrical conductivity.

When the sequence used hot air drying followed by an IR curing step, a final sheet

resistance value of 0.5 Ω/□ was achieved [11].

While the majority of conductive material research papers focus on silver, some

papers are starting to focus on copper screen paste. One such study was performed to

compare copper paste to silver paste in order to determine if there was any difference in

sintering performance between the two. The copper paste was cured on a hot plate at

150°C in an Argon atmosphere for 60 minutes. Argon shielding was necessary, since

curing of the copper in air resulted in poor conductivity due to the oxidation of uncoated

copper particles. The silver paste was cured in air also at 150°C for 15 minutes. The

copper paste produced a resistivity of 0.75 ∗ 10−4Ω ∗ cm, while the silver paste had a

resistivity of 0.87 ∗ 10−4Ω ∗ cm [8].The authors concluded that copper pastes were able

to provide comparable results to silver pastes.

An interesting related IR study of a copper screen paste actually suggested that

sintering copper pastes in an oxygen free environment was undesirable. The study

suggested that without oxygen, carbon based polymers, which are used to coat copper

particles, are not oxidized, meaning that they cannot reduce any oxide that is on the

particles. As a result one would see sooting, which are deposits on the surface on the

copper particles that can compromise quality of the sintering connections. An oxygen

free environment also means that pastes which contain glass frit, which is used as an

adhesion promoter, does not achieve the necessary bonding interface between the copper

print and the substrate. This results in poor adhesion and increased resistivity. One idea

12

drying atmosphere. This helped eliminate the sooting problem, and as a result, improved

adhesion and conductivity [12].

After reviewing a significant portion of the available literature on IR drying of

conductive inks and pastes, it is clear that most of the research has focused on silver inks

and pastes, since they can be sintered under the IR dryer. While there is some research

about IR drying of copper pastes, there is not a lot regarding copper inks. It is

understandable that since copper cannot be sintered under the IR dryer, that it would not

be researched as much as other inks that can be IR sintered. However, copper inks and

pastes still need to be dried in a high volume roll-to-roll printing process. IR drying is

compatible with such a process and should be explored further as a possible drying tool

for copper inks and pastes. In addition, it may be possible that use of an IR dryer for

copper inks and pastes might improve their sintering ability.

It was also apparent after reviewing the literature that there have been very few

studies done that use any sort of statistical analysis to determine the significant

parameters of an IR dryer used to dry any type of conductive material. While one study

used statistics and a mathematical model to determine the parameters for an IR dryer, it

was for drying a cotton web instead of conductive ink. It was found that drying of the

fabric was strongly affected by parameters such as the temperature of the dryer, the

dryer’s height, and speed at which the fabric moved through the dryer [13]. It may be

possible that these factors are significant in drying certain substrates that conductive ink

is printed on. The purpose of this particular study is use statistical analysis to evaluate the

significant parameters of an IR dryer used to dry copper ink and screen paste. This study

should help provide insights into a specific area of conductive ink research where there is

currently a gap in the literature.

This work focuses on exploring the possibility of using an infrared (IR) dryer to

process copper ink that is manufactured by Intrinsiq Materials for inkjet applications. The

objective is to determine which parameters on the IR dryer have a significant effect on

the drying of the inks and their performance on various substrates. Statistical analysis will

13

determine if the IR dryer will result in equal or better sintering performance when

compared with prints dried in a vacuum oven. If IR dried prints perform equally as well

as those that were oven dried, it will allow the IR dryer to replace the vacuum oven as the

preferred drying tool, thus significantly shortening the drying time, and allowing it to be

compatible with a high volume manufacturing process.

Currently, very few research papers have been written about using IR drying with

conductive inks. This is due in part to the fact that, as mentioned previously, conductive

inks are still a developing technology. Until recently, there has not been a strong focus on

using equipment that was compatible with a high volume manufacturing process. What

research has been on IR drying and conductive materials has mostly been done with

silver inks and pastes, with a few studies looking at copper screen paste. These studies

suggest that IR drying can be used successfully to sinter conductive materials, with

14

3

Methodology

3.1 Screening Experiments

Preliminary experiments were run in order to identify possibly significant IR

dryer parameters as well as sources of external variability that impact the resistance of

copper prints. Initial experiments were performed using an Afford-A-Flash Infrared dryer

from the Black Body Corporation. In some cases a Black Body Corporation Rheostat was

used to control the temperature of the IR dryer. Resistance performance of prints dried in

the Afford- A – Flash dryer, seen in Figure 10, was measured against benchmark oven

prints, which were dried using a VWR Scientific Products vacuum oven seen earlier in

Figure 4.

Figure 10: Afford-A-Flash infrared dryer

The basic process flow for these experiments began with the printing step. During

the printing step, conductivity bar traces were printed on PEL Paper, Kapton, or Glass

substrates with the same dimensions for each experiment. The pads on each end of the

trace were 3 millimeters by 3 millimeters, while the line connecting the pads was 1

millimeter wide, by 91 millimeters long. Each trace was between 1 and 2 microns thick.

Each print during these experiments was made using a 3 solvent ink formulation with a

1-15

methoxy 2-propanol. All samples were printed using a Dimatix DMP-2831 inkjet

materials printer. All but one of the initial experiments was done using a drop spacing of

20 μm, with one experiment using both 15 and 20 μm in order to study laydown

variability. All prints also consisted of conductivity bars that were printed horizontally on

the substrate. For all samples printed on Kapton and Glass, prior to printing, each piece

was cleaned with IPA and then dried using a heat gun.

After the printing step, all of the prints were dried in either the vacuum oven or

the IR dryer. For all experiments prints dried in the vacuum oven were dried at a

temperature of 50°C for a duration of 1 hour. Prints were IR dried at a wide variety of

times and heights, resulting in several different temperatures that differed from one

experiment to the next. These specific parameters are shown in more detail in Table 1.

All of the prints were then sintered using the Xenon Sintron 2000 system. Each of the

conductivity bars was exposed to a different sintering voltage, which controlled the

amount of UV light energy the trace was exposed to, and ranged from 460-1400 Joules.

After each conductivity bar was sintered, resistance measurements were taken from pad

to pad using a Fluke 77 IV Multimeter in order to determine the optimal sintering point

and to understand the size of the “sintering window.” An example of a print is shown in

Figure 11.

Figure 11: Inkjet print on kapton

The initial experiments were conducted to determine which IR dryer parameters

had a significant effect on resistance performance. The IR dryer parameters of dryer

height, dry time, dryer power, and the backing material used underneath the substrate

16

material, and the IR dryer was moved over top of the print at a pre-set height. Then using

a stopwatch, the desired dry time was measured. When the time had been reached, the

dryer was moved off the print. In some cases, the temperature of the dryer was decreased

using the rheostat. The rheostat was used to decrease the power to the IR dryer in order to

reduce temperature of dryer and to see if there was an advantage to drying at a lower

temperature for a longer time.

From the first few experiments, it was found that all four of these factors were

significant. However it was observed that there was a significant amount of

within-experiment variability. As a result, a fishbone diagram was created as shown in Figure 12

in order to brainstorm potential causes of the observed variability. It was decided that the

Machine, Method, and Materials sections should be the focus of further study, since they

were the most likely sources of variability. Additional preliminary experiments were

conducted to study time between drying and sintering, variability in the amount of ink

being put down by the Dimatix printer, and hot zones in the IR dryer, in order to

determine if any of those sources could be the cause of within experiment variability. A

summary of the experimental space explored using the Afford- A- Flash IR dryer can be

found in Table 1. The purpose of this table is to show the ranges of the various

17 Figure 12: Fishbone diagram to identify sources of within experiment variability

Table 1: Summary of initial experimental space explored

Dryer Height (inches)

Dry Time

(seconds) Substrate Rheostat (%)

Time between dry and sinter

Laydown Variability (drop spacing)

Hot

Zones? Backing Type Inks 10,20,40 PEL 30,40,45,60 Kapton 20,40 PEL 40,60 Kapton 20,40 PEL 40,60 Kapton 10,20,30,40 PEL 30,40,45,60,75,90,1 20, (3,4,5,6,7 mins) Kapton

3,4,5 minutes Glass 80 ~ 1 hour 20 µm NA Aluminum LAS-2013-5-45 Aluminum and Acrylic LAS-2012-02, LAS-2013-43 LAS-2012-02, LAS-2013-43 LAS-2012-02, LAS-2012-24, LAS-2013-43, LAS-2013-5-45 1 inch 2 inches 3 inches LAS-2012-02, LAS-2013-43 NA,100, 80 <30 mins, ~1 hour 20 µm NA

Aluminum Aluminum Parameters NA NA 1 hour 1 hour 20 µm Aluminum and Acrylic Co n d iti o n s 4 inches 20 µm 15µm,20 µm NA NA Yes NA,100, 80, 60, 40

18

In Table 1 where it says “NA” for hot zones it is important to note that hot zones

still existed at those conditions, but were not accounted for when running experiments.

Once it was discovered that hot zones were present they were only taken into account

when running experiments at a dryer height of 4 inches. No additional experiments were

run to study hot zones at lower dryer heights. It is also important to note that all the inks

listed in the table had the same solid loading fraction and 3 solvent formulation described

previously, with the main difference between each of the inks being the amount of

polymer coating on each of the copper particles.

3.2 Final Designed Experiment

While preliminary experiments were conducted using the Afford-A-Flash IR

drying system in order to demonstrate proof of concept, optimization of IR drying

parameters was done on a pilot high volume Vastex D-100 Infrared Curing System, seen

in Figure 13.

Figure 13: Vastex IR drying system

In order to be able to transfer what was learned from the Afford- A-Flash system

to the Vastex system, several characterization experiments were performed on the Vastex

system so that the robustness of initial conclusions could be tested. The first experiment

was designed to characterize belt speed setting as a function of the time the print

remained under the dryer. In order to characterize belt speed setting, pieces of glass were

placed on the conveyor belt, and a stopwatch was used to time how long it took the pieces

of glass to pass through the dryer for a given belt speed setting. The data shown below in

Belt Speed Setting Heat Setting

19

Figure 14 provides residence times for the corresponding belt speed settings. From this

data, it was concluded that small changes at lower belt speed settings resulted in a large

difference in residence time, with the opposite being the case at higher belt speed

settings.

Figure 14: Belt speed settings as a function of time

The second characterization experiment was designed to help determine the actual

temperature that different heat settings corresponded to. The belt speed settings and heat

settings for this experiment can be found in Table 2. All of these settings were tested at

both the 2 inch and 4 inch dryer heights.

Table 2: IR dryer parameter used in experiment to characterize heat setting

Belt Speed Setting: 5 10 15

Temperature:

4 1 print 1 print 1 print

Maximum Heat Setting

1 print 1 print 1 print

0 15 30 45 60 75 90 105 120 135 150 165 180 195

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105

Ti

m

e

(

sec

s)

Belt Speed Settings

20

In addition to testing these settings at different dryer heights, two different

substrates, Glass and Kapton, were used at both dryer heights to test each combination of

belt speed and heat setting. Using a Fluke 62 IR thermometer, temperature data was

collected on each piece of Kapton and Glass as they exited the dryer. The results of this

experiment can be found in Figures 15(a) through 15(d). From this data, it can be

concluded that substrates reach significantly higher temperatures at lower dryer heights

and slower belt speed settings than they do at higher heights and faster belt speed

settings. It appears that both belt speed and dryer height have a significant effect on the

exit temperature of the substrate.

(a) (b)

(c) (d)

Figure 15a,b,c,d: Substrate temperature for glass and kapton at specific dryer heights and heat settings

An additional experiment was run to better understand the relationship between

heat setting and temperature. After taping a thermocouple to a piece of Kapton with the

dryer height set to four inches, belt speed set to zero, and the heat setting set to its

21

dryer. The piece of Kapton was then placed under the dryer in all 4 corners and the

middle of the dryer for one minute, and the maximum temperature that was reached was

recorded. The Kapton was cooled to below 30°C in between each measurement. This

procedure was then repeated at a dryer height of two inches. Using the data from this

experiment, thermal heat maps were created at each of the two dryer heights to better

characterize the relationship between heat setting and temperature. These heat maps can

be seen in Figures 16(a) and 16(b).

(a)

(b)

Figure 16a,b: Thermal map at the highest heat setting at height of 2 (a) and 4 (b) inches

Once the characterization experiments for the Vastex system were completed,

several experiments were conducted to determine how the new IR system compared to

the vacuum oven, as well as what the optimal operating space was for the IR system. The

prints used during each of these experiments were of the same dimensions as the one

used in the preliminary experiments, shown previously in Figure 11. The process flow

for the prints in these experiments was identical to the one followed in preliminary

experiments, with a print going through the printing, drying, and sintering steps.

22

The purpose of the first experiment was to determine if a set of conditions similar

to the ones used on the Afford-A-Flash IR system could be used in the new Vastex IR

system, in order to match the resistance performance of oven dried prints. In addition to

the conditions displayed Tables 3 and 4, prints were also dried in the vacuum oven for an

hour at 50°C to be used for comparison to the IR prints.

Table 3: IR dryer parameters for prints dried at a height of 2 inches

Belt Speed Setting: 5 10 15

Temperature:

Minimum Heat Setting 1 print 1 print 1 print

4 1 print 1 print 1 print

Maximum Heat Setting 1 print 1 print 1 print

Table 4: IR dryer parameters for prints dried at a height of 4 inches

Belt Speed Setting: 5 10 15

Temperature:

4 1 print 1 print 1 print

Maximum Setting 1 print 1 print 1 print

Once it was demonstrated that the Vastex IR drying system was capable of

matching the performance of oven-dried prints, additional experiments were conducted in

order to identify the limit on the drying parameters for a specific ink and substrate

combination. The purpose of these experiments was to better define the operating space

for Intrinsiq’s ink so that the IR dryer parameters could be optimized through the use of a

designed experiment (DOE). Specifically, these experiments helped to determine the

maximum belt speed setting that the dryer could be run at by determining the setting at

which prints were no longer coming out of the dryer dry, even with the maximum energy

input. Maximum energy input was defined by setting the dryer at its lowest height and its

highest heat setting. The experiment then involved increasing the belt speed setting until

a setting was reached where the prints visibly were no longer being fully dried. By

defining the limits for each of the dryer parameters, it helped narrow the operating space

of the dryer and made designing the experiment for optimizing the dryer parameters

23

After completing the characterization and parameter limiting experiments, a final

experiment was designed to identify the optimal IR drying parameters for a specific

substrate and ink system. This experiment was conducted using a the same 3-solvent ink

formulation, of ethylene glycol, butonal, 1-methoxy 2-propanol with a 12% solid loading

fraction, and a fresh Dimatix 10 pico liter print cartridge. When transferring the ink from

its storage jar into the Dimatix print cartridge, it was stirred for 1 minute and then passed

through a 5-µm filter. All the prints for a specific height setting and specified number of

print pass were made on the same day with the same cartridge in order to eliminate

cartridge age or ink age as potential sources of variability in the experiment. A new

cartridge was then used the following day for a different height and number of printed

passes combination. Standard cartridge settings, meaning a cartridge temperature of 45°C

with a voltage setting of 34, and a drop spacing of 20μm were used for all the prints. All

prints were printed on the inside roll face of 5 mil thick Kapton, and were either single or

double pass. The time between printing and drying was held constant at 10 minutes for

each print.

Once the prints were dried, they were then sintered within a few hours using the

Xenon system with PFN 2 settings, meaning a UV light pulse 1 millisecond in duration,

on a white acrylic backing and a lamp height of 1.5 inches. Prior to the sintering of each

trace, the lamp was preflashed to ensure the prints were being exposed at the desired

sintering voltage. After each print was sintered, a lint free rag was used to wipe down the

lamp to ensure no blow off was stuck to the lamp which could cause scattering of the UV

sintering light leading to variable curing intensity on subsequent flashes. The Fluke

voltmeter was again used to measure the resistance of sintered traces, which was the

response variable for this experiment. Table 5 shows the conditions for each parameter

24 Table 5: Settings for IR dryer and Xenon sintering system used in the final experiment

Height (inches) Conveyor Speed (setting)

Heat (setting) Voltage (kv)

3 10,15, 20 5,6,High 2.7, 2.9, 3.1

4 10,15, 20 5,6,High 2.7, 2.9, 3.1

The prints were run and sintered in the following random order for both single

and double pass prints. Factor a= height, Factor b= speed, Factor c= heat, Factor d=

voltage. For Factors b,c, and d, lower case letters indicate the lowest parameter setting,

while uppercase indicates the middle setting, and uppercase and underlined indicates the

highest setting. For “A” lower case indicates a height of 3 inches, and uppercase indicates

a height of 4 inches.

1) aBcdDD = 3in, 15, 5, 2.7, 2.9, 3.1

2) abcdDD = 3in, 10, 5, 2.7, 2.9, 3.1

3) aBcdDD = 3in, 20, 5, 2.7, 2.9, 3.1

4) abCdDD = 3in, 10, 6, 2.7, 2.9, 3.1

5) aBCdDD = 3in, 15, 6, 2.7, 2.9, 3.1

6) aBCdDD = 3in, 20, 6, 2.7, 2.9, 3.1

7) aBCdDD = 3in, 20, High, 2.7, 2.9, 3.1

8) abCdDD = 3in, 10, High, 2.7, 2.9, 3.1

9) aBCdDD = 3in, 15,High, 2.7, 2.9, 3.1

The above conditions were completed a total of 3 times. Once the above conditions had

been repeated 3 times, the height was increased from three to four inches, meaning all “a”

s became “A” s, and then the cycle repeated again. Once the second cycle was complete,

25

4

Results

4.1 Feasibility of IR Drying

After the initial experiments were performed, a basic understanding of the

Afford-A-Flash IR Drying system had been achieved. The effects of dryer height, dry time and

dryer power on sintering performance had been evaluated. It was determined that these

parameters created a “drying continuum”, meaning that a similar amount of energy could

be put into prints at significantly different sets of conditions. For example, Figures 17 and

18 show the sintering curves for two different experiments. The x-axis of the two graphs

shows the voltage that was sent to the UV lamp to sinter each conductivity bar, and the

y-axis shows the resulting resistance of that bar. The main difference between the two

graphs is that Figure 17 shows results of the prints done using full dryer power, 1 inch

dryer height, and short dry times, while prints in Figure 18 were dried for much longer

times at a height of 4 inches, with decreased dryer power, which was controlled by a

rheostat. Despite the significant difference in conditions, both graphs show most

resistances falling between 17Ω and 30Ω. Therefore it can be concluded that a similar

amount of energy could be put into prints at significantly different sets of conditions.

Figure 17: Resistance values ranging between 17 and 30Ω were achieved using low dryer heights and short dry times

15 20 25 30 35

2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3

M

in

im

u

m

R

e

si

stan

ce

,

Ω

Voltage (kV)

26 Figure 18: Similar resistance values were achieved using high dryer heights, less dryer power and longer dry

times

Through these initial experiments it was also learned that it was much easier to

“over-dry” a print under the IR dryer than it is when drying a print in the vacuum oven.

An example of an over dried print can be seen in Figure 19. Figure 19 shows a series of 4

prints, all of which were dried at a height of 1 inch for 30, 45, 60, and 75 seconds,

respectively. While the sintered bars in the 30 and 45 second prints had a clear copper

color, the sintered traces in the 60 and 75 second prints have a much darker color. Figure

19 shows one of the unsintered traces that was dried for 75 seconds. There is a somewhat

bluish color, which is significantly different from the black color typically observed for

unsintered traces. Although not as obvious in the photo, this color change was also

observed in the 60 second dried print. Of the 4 prints, the best resistance observed was

7Ω, which was measured on the sample that was dried for 45 seconds. The best resistance number observed on the 60 second dried print was 28.2Ω, and an open circuit on the 75

second dried print. Between the resistance data and the observed color change, this

suggests that the 60 and 75 second dried prints were over-dried, which caused the copper

particles to oxidize giving the traces the blue color.

15 20 25 30 35

2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2

R

e

si

stan

ce

(

Ω

)

Voltage (kV)

27 Figure 19: Example of an over-dried IR print

After these preliminary experiments were conducted, it had been determined that

IR dried prints were comparable to prints that were oven dried. Data to support this can

be found in Figure 20. From the graph, one can conclude that IR dried samples had

slightly better sintering performance. Drop spacing refers to the distance between the

centers of each of the drops coming out of the Dimatix printer as the print is being made.

Figure 20: IR dried prints preformed slightly better than vacuum oven dried prints

In order to confirm these graphical conclusions, a hypothesis test was performed

to determine if the average resistance of IR dried traces and the average resistance of

oven dried traces were statistically different. Figure 21 shows that a p-value of less than

4.00 8.00 12.00 16.00 20.00 24.00 28.00

2.10 2.20 2.30 2.40 2.50 2.60 2.70 2.80 2.90 3.00 3.10 3.20 3.30 3.40

Re

si

stanc

e

(

Ω

)

Voltage (kV)

28

0.0001 was calculated for this test. That means that at α=0.05, it can be concluded that

the average resistance of 15.3 Ω for IR dried prints is statistically different from the

average resistance of 18.8 Ω for oven dried prints.

Figure 21: Hypothesis test comparing the average resistance of IR dried prints to oven dried prints

Although the IR dried prints performed as well as the vacuum oven dried prints,

there was a significant amount of variability within some of the initial experiments. As a

result, experimentation was done to try to identify and control sources of variability. The

first potential source of within-experiment variability studied was the backing material

used underneath the substrate being dried. This factor was studied because different

materials have different thermal conductivities, which means that some materials

dissipate heat faster than others. This may have a significant impact on the temperature

profile of copper during drying and sintering. The thermal conductivity of some common

materials can be found in Table 6. The two backing materials studied were aluminum and

29

0.250 inches thick. It was found that after a duration of 2 minutes under the IR dryer at a

height of 4 inches, the aluminum reached 355°C, while the acrylic only reached a

maximum temperature of 160°C. While, it took the aluminum only 30 seconds to cool

back to 23°C, it took the acrylic nearly one hour, which was not acceptable from a

repeatability standpoint. It was determined that the ideal backing material should be one

that does not get too hot to avoid damaging the substrate, as well as one that cools rapidly

so that it does not affect repeatability.

Table 6: Thermal conductivity properties of common materials

Another possible source of within-experiment variability that was analyzed was

whether or not the time between drying and sintering was a significant factor. The reason

this factor might cause variability is that prior to or during the drying process, the

polymer coating on the copper particles could degrade. Degradation of the polymer

coating would increase the probability of oxidation of the copper particles, which would

increase with time, and significantly affect conductivity. An experiment was conducted

which involved making 2 sets of prints and drying them under the same set of conditions

within the same time period. One of these sets was sintered right away, and the other was

sintered three hours later. Based on the data collected and displayed in Figures 22 and 23,

the time between drying and sintering did not have a significant impact on sintering

performance for the sintering delay tested. Therefore it can be concluded that time

between drying and sintering is not an important parameter provided the sintering is done

30 Figure 22: Resistance data for prints sintered within 30 minutes of drying

Figure 23: Resistance data for prints sintered 3 hours after drying

In order to confirm these graphical conclusions, a hypothesis test was performed

to determine if the average resistance of traces sintered immediately after drying and the

average resistance of traces sintered three hours after drying were statistically different.

Figure 24 shows that a p-value of 0.1254 was calculated for this test. That means that at

α=0.05 it can be concluded that there is no statistically significant difference between the

10 15 20 25 30 35 40 45

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2

R

e

si

stan

ce

,

Ω

Voltage (kV)

4 in 30 sec 4 in 45 sec 4 in 60 sec 4 in 90 sec 4 in 45 sec No RS Oven Check

10 15 20 25 30 35 40 45

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2

R

e

si

stan

ce

,

Ω

Voltage (kV)

31

average resistance of prints sintered immediately after drying and prints sintered three

hours after drying.

Figure 24: Hypothesis test comparing the average resistance of prints sintered immediately after drying to prints sintered three hours after drying

Experimentation was also done to determine if printer laydown variability, or

variability in drop spacing, was a possible source of within-experiment variability. Prior

to the experiment, all of the prints that had been analyzed were printed with a single pass

and a standard drop spacing of 20µm. A 20 µm drop spacing corresponds to 1270 dots

per inch, or DPI as it appears in the figures. It was hypothesized that this spacing may be

32

hypothesis, three prints were made at the standard drop spacing and three prints were

made at a drop spacing of 15µm (1693 dots per inch). All six prints were then dried under

the IR dryer and sintered so that resistance measurements could be taken. This same

experiment was repeated again with 6 prints being made, 3 at a drop spacing of 20μm and

3 at drop spacing of 15μm. The difference with the repeat experiment was that instead of

drying prints with the IR Dryer, prints were dried using the vacuum oven. Figures 25 and

26 below show a comparison of the IR and vacuum oven data collected from these two

experiments.

Figure 25: IR dried prints made with standard dots per inch (1270) and higher dots per inch (1693)

Figure 26: Oven dried prints made with standard dots per inch (1270) and higher dots per inch (1693)

From Figures 25 and 26 it can be concluded that there is definitely laydown

variability occurring regardless of whether the print is IR or oven dried. It was also

observed that the prints made with a drop spacing of 15µm out performed those with the

10 12 14 16 18 20 22 24

2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2

R e si stan ce ( Ω ) Voltage (kV)

3 mins STD DPI(1) 3 mins STD DPI(2) 3 mins STD DPI (3) 3 mins High DPI(1) 3 mins High DPI(2) 3 mins High DPI (3)

10 12 14 16 18 20 22 24

2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3

R e si stan ce ( Ω ) Voltage (kV)

33

standard drop spacing of 20µm in both drying systems. Figures 27 and 28 show range

data that was calculated for each drop spacing condition at each sintering voltage. Since 3

prints were made at each drop spacing condition and sintered at the same set of voltages,

the difference between the highest and lowest resistance values at each sintering voltage

were calculated in order to quantify variability. This was done for both the prints dried in

the vacuum oven as well as the prints dried using the IR Dryer.

Figure 27: Resistance range for IR dried prints with two different dots per inch settings

0 1 2 3 4 5 6

2.3 2.5 2.7 2.9 3.1 3.3

R

e

si

stan

ce

r

an

ge

, Ω

Voltage, kV

34 Figure 28: Resistance range for oven dried prints with two different dots per inch settings

Figures 27 and 28 show the resistance range of the 3 prints, or the variability at

each specific drop spacing condition, in both the vacuum oven and IR systems. The

graphs also shows the range, or variability, between the two drop spacing conditions

which is represented in the figures by the “Both DPI” series. From this data, it was

determined that near the optimum sintering voltage, which is roughly 2.3 or 2.4 kilovolts

based on Figures 25 and 26, the oven is more reproducible than the IR dryer. This was

determined by comparing the range values 2.3 and 2.4 in Figure 28 to the same values in

Figure 27. In the oven system, the range values at those voltages fall between zero and

three, but in the IR system, the range values fall between two and five at the same

voltages. However, once moving away from the optimum sintering voltage, the oven’s

reproducibility begins to decrease significantly, while it continues to stay constant for the

IR dryer. Therefore the IR dryer offers wider process latitude, and as long it is controlled

is not a source of within experiment variability.

The last factor that was studied as a potential source of within experiment

variability was the possibility of hot zones in the IR dryer. Initially, a simple test was

done to determine if any hot zones existed under the Afford- A- Flash IR dryer as well as

their location. This test involved taping off the perimeter of the IR dryer, and placing

0 1 2 3 4 5 6 7 8

2.2 2.4 2.6 2.8 3 3.2

R

e

si

stan

ce

r

an

ge

, Ω

Voltage, kV

35

white photo paper underneath the dryer for an extended period of time so that the paper

would begin to change color. A picture of this test can be seen in Figure 29.

Figure 29: The temperature in the center of the IR dryer is higher since it caused more discoloration

Figure 29 clearly shows that the center of the paper was discolored much more

than other areas of the dryer, implying that the temperature in the center of the dryer was

higher than it was on the outer edges of the dryer. This indicated that hot zones do exist,

are a source of variability, and need to be taken into account when running IR

experiments.

While initial experiments proved that IR dried prints could match the performance

of oven dried prints, follow up experiments indicated that there are some sources of

within-experiment variability that can affect the resistance results of each experiment.

Several of these factors were explored, and their impact on resistance performance

evaluated so that they could be controlled in future experiments. While there are still

other possible sources of external variability that could be evaluated, the decision was

made to switch drying systems to a Vastex dryer, which is a pilot roll-to-roll system, to

try to eliminate some of the external sources of variability that were present in the

36 4.2 Characterization of Vastex Dryer

Prior to optimizing the dryer parameters, an experiment had to be conducted in

order to determine if prints dried in the pilot high volume Vastex system could match the

conductivity performance of those dried in the vacuum oven. Once the characterization

experiments discussed in the methodology section were completed, this comparison

experiment was executed. Single pass prints were made on Kapton, and data was

collected at 2 different dryer heights. The results of the experiment can be seen in Figures

30 and 31.

Figure 30: Resistance data for IR and oven dried prints at 2 inches

Figure 31: Resistance data for IR dried and oven dried prints at 4 inches

15 20 25 30 35 40 45 50 55 60

1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3 3.4

R e si stan ce ( Ω ) Voltage (kV)

Resistance Oven Dry Resistance IR-Heat:low Belt:15 Resistance IR-Heat:low Belt:10 Resistance IR-Heat:low Belt:5 Resistance IR-Heat:4 Belt:15 Resistance IR-Heat:4 Belt:10 Resistance IR-Heat:4 Belt:5 Resistance IR-Heat:high Belt:15 Resistance IR-Heat:high Belt:10

Resistance IR-Heat:high Belt:5 Best Oven (1) Best Oven (2)

10 14 18 22 26 30 34 38 42

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3 3.4

R e si stan ce ( Ω ) Voltage (kV)

Oven Dry IR-Heat:4 Belt:15 IR-Heat:4 Belt:10 IR-Heat:4 Belt:5 IR-Heat:high Belt:15

37

From Figures 30 and 31, it can be concluded that it is possible for the Vastex IR

dryer to match the resistance performance of vacuum oven dried prints. In Figure 30, we

see that the sintering curves for all IR prints dried at a heat setting of four, regardless of

their belt speed setting, matched the sintering curves for the oven dried prints.

Additionally, the print dried at the high heat setting with a belt speed setting of 15 also

matched the sintering curves for the oven dried prints. We also see that none of the prints

dried at the low heat setting or at the high heat setting and slower belt speed settings

performed as well as the oven. In Figure 31, we see that all IR dried prints at the heat

setting of four performed just as well or better than the oven dried prints. Additionally,

the performance of the high heat drying test with a belt speed setting of ten is similar to

the oven with the other two high heat prints having slightly worse performance. This data

indicates that under these conditions, the Vastex system is capable of drying prints, which

produce comparable sintering performance as prints dried in the oven.

In order to confirm these graphical conclusions, a hypothesis test was performed

using the data from Figures 30 and 31 to determine if the average resistance of IR and

oven dried traces were statistically different. Figures 32 and 33 show that a p-value of

0.0008 and 0.0078 were calculated for this test, respectively. That means that at α=0.05,

it can be concluded that there is a statistically significant difference between the average

38 Figure 32: Hypothesis test for 1 pass prints dried at 2 inches with the new IR dryer and the oven

39

A follow-up experiment was done to determine how double printing pass IR dried

prints would perform with the Vastex IR drying system in comparison to the vacuum

oven. The results of this experiment can be seen below in Figure 34.

Figure 34: Resistance data for double pass IR and oven dried prints at 4 inches

From Figure 34, we can see that prints dried with all three of the heat settings and

belt speed settings of 10 were the closest to matching the performance of the oven dried

print. This is different from what we saw with the single pass prints, where the belt speed

setting of 15 was the closest to matching the performance of oven dried prints. This is

because the dryer now needs to dry a thicker trace, which means it needs to move at

slower speed so it has more time to dry the traces.

In order to confirm these graphical conclusions a hypothesis test was performed to

determine if the average resistance of IR dried traces and the average resistance of oven

dried traces were statistically different. Figure 35 shows that a p-value of 0.0544 was

calculated for this test. That means that at α= 0.05, it can be concluded that the average

resistance for IR dried prints is not statistically different from the average resistance for

oven dried prints.

10 15 20 25 30 35 40

2.20 2.30 2.40 2.50 2.60 2.70 2.80 2.90 3.00 3.10 3.20 3.30 3.40

R

e

si

stan

ce

(

Ω

)

Voltage (kV)